Retinal Degeneration in tulp1 ^−/− Mice: Vesicular Accumulation in the Interphotoreceptor Matrix

Abstract

purpose. The Tulp1 gene is a member of the tubby gene family with unknown function. Mutations in the human TULP1 gene cause autosomal recessive retinitis pigmentosa. To understand the pathogenic mechanism associated with TULP1 mutations and to explore the physiologic function of this protein, we examined tissue distribution of the Tulp1 protein in normal mice and the photoreceptor disease phenotype in Tulp1–ablated mice.

methods. Tissue distribution of the Tulp1 protein in normal mice was examined by immunoblotting and immunocytochemistry. The disease phenotype in tulp1 ^−/− mice was studied by light and electron microscopy, electroretinography (ERG), and immunocytochemistry. These results were compared with another mouse model of retinal degeneration carrying a rhodopsin mutation.

results. Tulp1 is found exclusively in photoreceptors, localizing predominantly in the inner segments. It is a soluble protein with an apparent molecular weight of approximately 70 kDa. Photoreceptor degeneration developed in tulp1 ^−/− mice, with early involvement of both rods and cones. At the early stage of degeneration, rod and cone opsins, but not peripherin/RDS, exhibited prominent ectopic localization. Electron microscopy revealed massive accumulation of extracellular vesicles surrounding the distal inner segments.

conclusions. The function of Tulp1 is required to maintain viability of rod and cone photoreceptors. Extracellular vesicular accumulation is not a common phenomenon associated with photoreceptor degeneration but appears to be a distinct ultrastructural feature shared by a small group of retinal disease models. The defect in tulp1 ^−/− mice may be consistent with a loss of polarized transport of nascent opsin to the outer segments.

The tubby gene family has at least four members that are defined by the highly conserved carboxyl-terminal half of their primary sequences. ^{1 2} These include tubby, the prototype of this family, and Tulp1, Tulp2, and Tulp3, for tubby-like proteins 1, 2, and 3, respectively. The physiological functions of any member of this family are not known, but two have been linked to neurosensory disorders. A recessive mutation in the tubby gene causes photoreceptor and cochlear degeneration, and adult-onset obesity in mice. ^{3 4 5 6} Recently, genetic mutations in the human TULP1 gene were found to cause autosomal recessive retinitis pigmentosa (RP). ^{7 8 9} Most of these mutations reside in the conserved C-terminal half of the coding sequence. 

RP is a genetically and phenotypically heterogeneous group of inherited retinal diseases that lead to the eventual degeneration of the rod and cone photoreceptor cells. Early stages of typical RP are characterized by progressive night blindness and loss of midperipheral visual field. Later in the disease, patients lose far peripheral field and central vision as well. Clinically, most RP patients have a primary rod defect and later involvement of the cones. It is believed that in these cases the cone defect is secondary to the loss of rods. A large number of gene mutations have been identified as causes of RP in the past decade (http://www.sph.uth.tmc.edu/RetNet). Most of these genes encode proteins that reside in the photoreceptor outer segments and participate in the rod phototransduction cascade, which begins with photon capture and leads to membrane hyperpolarization and neural signaling. Visual transduction has been intensely investigated for years. Functions of the components in this pathway are therefore known, and the consequences of their defects are often understood. More recently, a number of RP-related genes have been identified that do not appear to participate in phototransduction but that have functions that are otherwise unknown. Among this group of genes is Tulp1. 

To begin to understand the physiological function of Tulp1, we have determined the normal protein distribution and analyzed the retinal disease in tulp1 ^−/− mice in comparison with another mouse model of retinal degeneration due to a dominant rhodopsin mutation affecting primarily rod photoreceptors. We describe the distinct histopathologic features in the tulp1 ^−/− retina. 

Methods

Animals and Genotyping Analysis

Heterozygous tulp1 knockout mice, in which the coding sequence was disrupted by the insertion of a neomycin-selectable marker, were generated at AxyS Pharmaceuticals (La Jolla, CA; M. North and M. Duyao, unpublished data). Homozygotes (tulp1 ^−/−) were derived from heterozygous mating and identified by PCR amplification of genomic DNA using primer pairs in exon 8 of the Tulp1 sequence: 5′-AAGGAGGAGAGAGCCTCTTC, sense, and 5′-TTCTCAGTGTCCAGGTGCAG, antisense; and a pair of primers for the neo^R sequence, 5′-ACAATCGGCTGCTCTGATGC and 5′-GTCACGACGAGATC-ATCGC. All four primers were used together in a 20-μl polymerase chain reaction (PCR) buffer (pH 8.6) containing approximately 100 ng DNA, 20 picomoles of each primer, 200 μM each dNTP, 1.5 mM MgCl₂, and 10% dimethyl sulfoxide. PCR reactions were cycled 35 times at an annealing temperature of 55°C and an elongation temperature of 72°C. A 167-bp product and a 450-bp product identified the wild-type (wt) and the knockout alleles, respectively. A transgenic mouse line carrying a dominant rhodopsin mutation T17M was created by pronuclear injection of a DNA construct containing the full length of human rhodopsin gene with 4.8 kb of 5′ and 6.2 kb of 3′ flanking sequences and an engineered threonine-to-methionine mutation in its coding sequence. This line of mice had been used in a previous study of therapeutic modalities for retinal degeneration ¹⁰ and was used in this study to compare the degeneration phenotypes between the two genetically distinct mutants. The Fvb/n mice homozygous for the rd allele were obtained from the Charles River Laboratory (Wilmington, MA). The rd allele carries a nonsense mutation in the gene encoding the β-subunit of cGMP phosphodiesterase. ¹¹ Homozygous rds mice in the Balb/c background were obtained from Maisy Tang at the New England Primate Research Center (Southborough, MA). The rds allele carries a mutation in the gene encoding the peripherin/RDS protein, a structural protein located in the outer segment disc membrane. ^{12 13} All experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Immunoblot Analysis and Immunocytochemistry

Two antibodies specific for the Tulp1 protein were used in this study. A 17-mer synthetic peptide was made and used to generate peptide antibodies in rabbits, which were then affinity purified (Quality Controlled Biochemicals, Hopkinton, MA). The peptide sequence KKPETPDSLESKPRKAG, corresponding to amino acid residues 44–60 of the mouse Tulp1 sequence (GenBank accession AF085681), was chosen based on its divergence from other members of the tubby gene family, as well as on predicted surface probability and antigenicity. An antibody against a fusion protein corresponding to the divergent N-terminal half of the human TULP1 protein was raised in rats (AxyS Pharmaceuticals ¹ ) and used as crude serum. For immunoblot analysis, isolated retinas and other somatic tissues were solubilized in Laemmli buffer, separated on 10% sodium dodecyl sulfate–polyacrylamide gels, transferred to polyvinylidene difluoride membranes, incubated with primary antibodies and peroxidase-conjugated secondary antibodies, and detected using chemiluminescent reagents (Pierce, Rockford, IL). An anti-actin antibody was used to probe the same blots as approximate loading controls (Cat. No. A-2066; Sigma, St. Louis, MO). To determine whether the Tulp1 protein is present in the cytosol or whether it exhibits membrane-cytoskeleton association, isolated retinas were disrupted by one cycle of freeze-thaw; suspended in a buffer containing 50 mM Tris [pH 7.5 ], 150 mM NaCl, 4 mM MgCl₂; and centrifuged at 160,000g for 1 hour at 4°C. Supernatant and pellet were then subjected to immunoblot analysis as above. For immunocytochemistry, mouse eyes were fixed in 4% formaldehyde in phosphate-buffered saline for 1 hour. The anterior segments and lens were removed. The eyecups were cryoprotected in 30% sucrose, shock frozen in liquid nitrogen, and sectioned at 10-μm thickness. All sections were cut along the superior to inferior axis so that the density gradient of cones could be visualized along this axis. ¹⁴ Sections were incubated with primary antibodies against Tulp1, rod opsin (rho-4D2 for the N terminus and rho-1D4 for the C terminus, ¹⁵ ) blue (short-wavelength sensitive) and green (middle-wavelength sensitive) cone opsins, ¹⁶ and peripherin/RDS. ¹⁷ This was followed by incubation with Cy3-conjugated secondary antibodies. Sections were viewed and photographed under a 40× objective on an epifluorescence microscope. Higher magnification images were acquired under a 100× objective on a confocal laser scanning microscope (TCS4D; Leica, Deerfield, IL). 

Electroretinography

Corneal electroretinograms (ERGs) were recorded from dark- and light-adapted mice as previously described. ¹⁰ Dark-adapted ERGs measure rod function in response to a flash of light. Light-adapted ERGs, performed in the presence of background lighting, isolate the cone responses. ¹⁸  

Light and Electron Microscopy

Microscopic examination of mouse retinas were performed essentially as described previously. ¹⁰  

Results

To determine the tissue distribution of Tulp1, immunoblotting analyses were performed using both Tulp1 antibodies on a number of different tissues. As shown in Figure 1 (upper left), only retina displayed a single band migrating at an apparent molecular weight of approximately 70 kDa, which is slightly larger than the calculated molecular weight for the mouse Tulp1 polypeptide (∼60 kDa). Both the peptide antibodies and the fusion protein antibodies recognized a band of the same molecular weight on immunoblots. Thus, Tulp1 expression was restricted to the retina. To further localize Tulp1 within the retina, retinal homogenates from wt and adult rd/rd mice were probed together on immunoblots with the Tulp1 antibodies. In adult rd/rd retinas, the photoreceptor layer, but not the other layers of the neural retina, had degenerated to completion. As shown in Figure 1 (upper right), the Tulp1 signal was absent in this photoreceptorless retina, indicating that Tulp1 is expressed specifically in photoreceptor cells. The Tulp1 signal was also greatly diminished in 2-month old rds (retinal degeneration slow) mouse retinas ¹² in line with the extent of photoreceptor cell loss in these retinas (data not shown). Retinal homogenate from a 3-week-old tulp1 ^−/− mouse, which had retained most of its photoreceptors (see below), was found not to contain Tulp1, thereby confirming the disrupted expression of Tulp1 in these mice. After a high-speed centrifugation of retinal homogenates from wt mice, to pellet all membranous and cytoskeletal fractions, Tulp1 was found exclusively in the supernatant (Fig. 1 , lower left). Thus, there was no stable association between Tulp1 with any particulate fractions under our conditions of homogenization. 

The photoreceptor-specific expression of Tulp1 and the absence of Tulp1 in the tulp1 ^−/− photoreceptors were further confirmed by immunocytochemical staining of retinal sections (Fig. 2) , using both Tulp1 antibodies. These data show that Tulp1 was distributed predominantly in the inner segments of the photoreceptor cells of wt retinas. To a much lesser extent, there was also staining in the perinuclear and synaptic regions (Fig. 2 , left and middle; wt retina at 3 weeks of age). The outer segments showed no Tulp1 staining. An age-matched tulp1 ^−/− retina showed no expression of Tulp1 in the photoreceptor layer or any retinal layers (Fig. 2 , right). The staining pattern in older wt retinas was the same as that at 3 weeks of age (not shown). 

Mutant mice lacking Tulp1 expression showed a progressive retinal degeneration that was already apparent by 3 weeks (Fig. 3) . At this age, the outer and inner segments were shortened, and the outer segments were disorganized. The outer nuclear layer was close to normal thickness but contained frequent pyknotic nuclei indicating ongoing photoreceptor cell death. The retinal pathology became progressively severe and appeared to reach end stage before 5 months of age. Retinal function, revealed by ERG (Fig. 4) , also declined with age in parallel to the deterioration of retinal morphology. Both rod and cone ERG amplitudes were markedly reduced at 4 weeks, and cone ERG became unrecordable by 8 weeks of age. Heterozygous animals as old as 4 months of age retained normal retinal function as shown by ERG testing (Fig. 4) . 

To follow the histopathologic changes in cone photoreceptors, we stained tulp1 ^−/− retinas ranging from 17 days to 4 weeks (not shown) of age with antibodies specific for blue and green cone opsins (Fig. 5) . These were compared with the T17M rhodopsin mutant and wt retinas (Figs. 5A 5B 5C 5D 5E 5F) . A similar pathologic condition was observed between the blue and green cones in tulp1 ^−/− retinas. There was ectopic staining of the inner segment and nuclear and synaptic layers by the cone opsin antibodies. There was only rudimentary outer segment formation at 17 days of age, the earliest time point examined (Figs. 5A , 5D ). At 4 weeks, cones became sparse in the tulp1 ^−/− retinas (not shown). In contrast, the T17M retinas had nearly normal-appearing cones at 3 weeks (not shown). In the T17M retinas, even at 7 weeks when a substantial number of rods had been lost, both blue and green cone opsins still exhibited a predominant outer segment localization, and the cones were present in approximately normal densities (Figs. 5B , 5E ). This is consistent with cones being affected secondary to rod disease in the rhodopsin mutant retinas. It appears that cones were affected earlier and more severely in tulp1 ^−/− mice. Thus, Tulp1 function may be essential in both rods and cones. 

Staining for rod opsin (Figs. 5G , 5H , 5I ) showed similar signal intensities in the outer and inner segments and prominent staining in the nuclear layer of the tulp1 ^−/− retina, both at 17 days and at 4 weeks of age (not shown). Thus, at the level of light microscopy the predominant outer segment staining pattern for opsin seen in wt retinas was lost in the tulp1 ^−/− retina. In comparison, the T17M retinas at an early stage of degeneration (4 weeks) preserved predominant outer segment localization of rod opsin, with only slightly increased staining of the inner segments and nuclear layer compared with a wt retina. Therefore, ectopic localization of both rod and cone opsins was an early hallmark of retinal disease in the tulp1 ^−/− retinas. Interestingly, staining for another outer segment–specific membrane protein, peripherin/RDS, showed partitioning of this protein in the outer segments of the tulp1 ^−/− retinas, similar to that found in the T17M or the wt retinas (Figs. 5J , 5K , 5L ). 

Electron microscopy revealed that the tulp1 ^−/− retinas had disorganized and shortened outer segments and shortened inner segments, consistent with the findings by light microscopy. At 3 weeks of age, there was a massive accumulation of vesicular profiles in the interphotoreceptor matrix (Fig. 6A ). The highest abundance was around the ellipsoid (distal) inner segments, tapering off toward the proximal inner segments and the outer segments. The vesicles were relatively uniform in size averaging 0.1 to 0.2 μm in diameter and bounded by a single membrane. No such vesicles were observed in the degenerating T17M retina or the wt retina (Figs. 6B , 6C ). Vesicle accumulation quickly diminished after peaking at 3 weeks of age, and few vesicles were observed in tulp1 ^−/− retinas at 5 or 8 weeks (data not shown). The extracellular location of the vesicles could be better visualized in tangential sections through distal inner segments (Fig. 6D) . Packets of vesicles were found clearly in the extracellular space among the profiles of several distal inner segments (as indicated by the abundance of mitochondria within them). Cross-sectional views of connecting cilia (Fig. 6D) had the normal 9 + 0 arrangement of microtubule pairs. Scanning through a number of tangential sections did not reveal any abnormal connecting cilia (from ∼50 examined). At higher magnification (Fig. 7) , the inner segment plasma membranes were marked by numerous protrusions (blebbing), some of which appeared to be in the process of pinching off. It is not yet clear whether these plasma membrane protrusions are in fact sites of origin of the accumulated vesicles, because membrane blebbing often accompanies apoptotic cell death. There was also an abundance of cytoplasmic vacuoles of varying sizes in the distal inner segments of tulp1 ^−/− photoreceptors. 

Discussion

Tulp1 is a photoreceptor protein of unknown function. Search of sequence databases did not reveal any significant homology with known proteins or functional motifs, other than members of the tubby family. Early-onset progressive retinal degeneration develops in mice deficient in Tulp1, affecting both rods and cones. In typical retinal degenerations, cone photoreceptors eventually die, even though the primary disease may affect only rods, because of an undefined secondary effect. In tulp1 ^−/− mice, there is primary involvement of both rods and cones. This is indicated by the abnormal cone opsin distribution, a near absence of cone outer segments, and diminished cone ERG responses in young animals. The loss of cone ERG responses occurred much earlier than that observed in mice carrying the T17M rhodopsin mutation (unpublished data) or a P23H rhodopsin mutation ¹⁹ ; both mutations primarily affect the rods only. By analogy with the tulp1 ^−/− murine model, involvement of cone photoreceptor cells is likely also to be a part of the primary retinal disease in human patients with RP due to TULP1 mutations. These patients manifest an early-onset retinal disease and have a more severe visual handicap than the average patient with RP due to other genetic defects. By their early thirties, the patients’ rod ERG responses are undetectable, and cone ERG responses are minimal. ⁷ This explains the severity of visual deficit, because only cones are essential for visual acuity and daytime vision. 

A marked feature in the tulp1 ^−/− retina is the massive accumulation of vesicles in the interphotoreceptor matrix surrounding the distal inner segments. We do not believe these to be debris from degenerating photoreceptors. That would imply that accumulation of extracellular vesicles is an expected feature in models of photoreceptor degeneration as long as the severity of disease is at or above a certain level. In fact, many rodent models of photoreceptor degeneration do not exhibit this phenotype. ^{20 21 22 23} Vesicular accumulation was first described in the rds mice in which vesicular elements amass in the subretinal space just distal to the rudimentary connecting cilia. ^{24 25} These vesicular profiles are immunopositive for opsin. ^{26 27 28} The rds photoreceptors never elaborate outer segments, because of the absence of a disc membrane structural protein, peripherin/RDS. ²⁹ Therefore, the vesicular elements in the rds retinas are likely to be the product of aborted outer segment formation. Treatment of isolated Xenopus retinas with tunicamycin also induced vesicular accumulation at the extracellular space between the rod inner and outer segments, accompanied by an arrest of new disc formation. ³⁰ These findings may be explained by the suggestion that nonglycosylated opsin and/or other membrane proteins are incompetent to support disc morphogenesis. Thus, in both the rds and the tunicamycin-induced models, it appears that nascent disc-building materials could be targeted to the approximately correct location, but disc morphogenesis could not be completed. Three other genetic murine models of retinal degeneration, in which outer segment formation occurs, have also been reported to accumulate extracellular vesicles accompanying photoreceptor degeneration. The pcd mice have photoreceptor and cerebellar Purkinje cell degeneration ^{31 32} ; the identity of the pcd gene is not yet known. The tubby mice, equivalent to rd5, have both retinal and auditory defects. ⁵ A point mutation at the C terminus of rhodopsin, P347S, results in a dominant photoreceptor degeneration in transgenic mice. ³³ All three exhibit extracellular vesicle accumulation in the interphotoreceptor matrix surrounding the distal inner segments that appear similar in size, density, and location to that seen in tulp1 ^−/− mice. Thus, the available data indicate that vesicular accumulation is a distinct phenotype attributable to a subset of gene mutations that cause photoreceptor degeneration. 

In the vertebrate retina, disc membranes in the photoreceptors are continually renewed by the addition of new membranes at the base and the loss of packets of older discs from the tip. Rhodopsin, the most abundant membrane protein in the outer segments, and other outer segment-bound proteins must be continually sorted and trafficked through transport vesicles from the trans-Golgi network toward the outer segments. ^{34 35} This is an essential, but poorly understood, aspect of photoreceptor cell biology. ³⁶ A number of small guanosine triphosphatases appear to play a prominent role. ³⁷ Polarized protein trafficking in general requires a large number of cytoplasmic components as well as interactions with signal sequences present on the cargo proteins. ³⁸ A single defect in these complex interactions could result in a loss of polarity in protein trafficking. There is biochemical evidence that the C terminus of rhodopsin may serve as a signal sequence in this process. ^{39 40} In transgenic mice, a mutation affecting the conserved penultimate proline residue (P347S) of rhodopsin causes a vesicular accumulation phenotype, ³³ the severity of which varies with levels of transgene expression. This is in contrast to mice bearing rhodopsin mutations not known to affect the trafficking signal and that do not exhibit this phenotype. ^{10 21 23} It can be argued that in the tulp1 ^−/− mice, as well as in the P347S mice, the accumulated vesicles may represent misrouted transport carriers for opsin. In support of this notion, both rod and cone opsins were ectopically localized early in the tulp1 ^−/− photoreceptors, suggesting that it is part of the primary defect in the tulp1 ^−/− photoreceptors. The vesicle accumulation phenotype in tulp1 ^−/− mice may be consistent with a loss of polarized trafficking of certain outer segment-bound membrane proteins. We speculate that in the tulp1-/-, tubby, P347S, and pcd mice, the mutations affect functions that regulate the polarity of normally outer segment–bound vesicular traffic. This results in a reduced number of vesicles reaching the correct target membranes; therefore, the outer segments are shortened, even though the nascent membrane may not be defective in its disc morphogenesis potential. Ectopically targeted vesicles may bud from lateral inner segment plasma membranes and accumulate in the interphotoreceptor matrix. This hypothesis implies that tubby, Tulp1, and pcd genes encode essential functions in this pathway. Interestingly, the RDS protein has a normal localization in the tulp1 ^−/− photoreceptors, supporting the notion that rhodopsin and the RDS protein are sorted to different transport vesicles, as noted previously. ⁴¹ It is also interesting to note that two members of the tubby gene family are now correlated with this phenotype, suggesting that this gene family may participate in polarized protein trafficking in different cell types. In-depth investigation into their functions and those of any newly found genotypes that produce the distinct phenotype discussed earlier may provide a genetic dissection of the process of polarized trafficking in photoreceptor cells. It also promises to broaden our understanding of the pathogenesis of RP and allied retinal degenerative diseases. 

 Supported by National Institutes of Health Grant EY10309, the Foundation Fighting Blindness, and the Massachusetts Lions Eye Research Fund. SAH is the recipient of postdoctoral training Grant EY06915. TL is the recipient of a career development award from Research to Prevent Blindness.

 Submitted for publication April 7, 1999; revised June 28, 1999; accepted July 14, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Immunoblot analyses of Tulp1 expression using Tulp1 antibodies. Actin was included to control for loading equivalency. Top left: multiple tissue blot showing specific expression in retina. Top right: retinal homogenates from mice of different genotypes. Tulp1 is detected in wt retina but not in photoreceptorless (rd/rd) or tulp1 ^−/− retinas. Lower left: Tulp1 is present in a high speed supernatant (S) but not in the pellet (P) fraction of retinal homogenates.

Figure 2.

 

Localization of Tulp1 by immunofluorescence. Left: wt retina at 3 weeks of age. Staining is most prominent in the inner segments. Middle: higher magnification of the same wt retina focusing on the inner segment layer. Right: tulp1 ^−/− retina at 3 weeks of age (littermate of the wt mouse shown in the left and middle panels) showing no significant staining in the photoreceptor layer by the Tulp1 antibodies. Staining of vascular profiles (seen in the choroidal layer just above the retinal pigment epithelium and as bright dots in the inner neural retina) was due to cross-reactivity to the circulating mouse IgG with the secondary antibody (goat anti-rat). IS, inner segment; OS, outer segment; ONL, outer nuclear layer.

Figure 3.

 

Photoreceptor degeneration shown by light microscopy. Shortened and disorganized photoreceptor outer segments were apparent at 3 weeks of age, and degeneration was complete by 5 months. RPE, retinal pigment epithelium; OS, outer segment; IS inner segment; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 4.

 

ERGs of tulp1 ^−/− and tulp1 ^+/− mice. Dark-adapted ERGs measure rod function, whereas light-adapted ERGs isolate the cone response. From top to bottom: a tulp1 ^+/− mouse at 4 months of age exhibited normal waveforms and amplitudes for rods and cones. The rod and cone ERG amplitudes of tulp1 ^−/− mice were severely diminished at 4 and 8 weeks and declined to an undetectable level by 5 months.

Figure 5.

 

Immunofluorescent localization of cone opsins, rod opsin, and peripherin/RDS in tulp1 ^−/−, T17M, and wt retinas. The genotypes of mice are indicated at the left and the target proteins recognized by the antibodies are marked at the top. (a) Ectopic localization of blue cone opsin in inner segments, nuclei, and synapses of tulp1 ^−/− photoreceptors; (d) similar staining pattern for green cone opsin in tulp1 ^−/− photoreceptors as in (a); (g) ectopic localization of rod opsin in tulp1 ^−/− photoreceptors, indicated by equal staining intensities in outer and inner segments and heavy staining of the nuclear layer. Shown are sections stained with rho-4D2, which are identical with those stained with rho-1D4 (not shown); (j) normal outer segment localization of RDS protein in tulp1 ^−/− photoreceptors. The ages of the animals were: tulp1 ^−/−, 17 days; T17M (b, e), 7 weeks; T17M (h, k), 4 weeks; and wt, 6 weeks. The tulp1 ^−/− mice were shown at the earlier age for best photoreceptor preservation. Photoreceptor outer segments and cell bodies will degenerate further at older ages. Older T17M retinas were shown to illustrate the differences in cone involvement in these two mutants.

Figure 6.

 

A distinct ultrastructural feature of tulp1 ^−/− photoreceptors revealed by electron microscopy. (A, B, C) longitudinal sections; (D) tangential section. (A) Three-week-old tulp1 ^−/− retina showing the region between the distal inner segment and basal outer segment. Three packets of extracellular vesicles (∗) are seen between adjacent inner segments. (B) Degenerating T17M retina is completely free of the extracellular vesicles. (C) Three-week-old wt retina. (D) Tangential section of a tulp1 ^−/− retina at the level of distal inner segments. Packets of accumulated vesicles (∗) are seen in the interphotoreceptor matrix between adjacent inner segments. Three connecting cilia (arrowheads) have the typical 9 + 0 arrangement of microtubules. M, mitochondria; IS, inner segment. Bar (A, B, C) 1 μm; (D) 0.4μ m.

Figure 7.

 

Higher magnification electron micrographs of 3-week-old tulp1 ^−/− photoreceptors. (A) Longitudinal section showing the basal outer segment (OS) and distal inner segment (IS) of one photoreceptor cell (left), connecting cilium (CC) of a neighboring photoreceptor, and a collection of extracellular vesicles surrounding the distal inner segment. Protrusions from the inner segment plasma membrane (arrowheads) were frequently observed. (B) Tangential section through distal inner segments. Plasma membrane protrusions are marked by arrowheads. Low-density intracellular vesicles (∗) bounded by a single membrane were also present in this region. M, mitochondria. Bar, 0.2 μm.

The authors thank Gary Yue for expert technical assistance, Robert Molday for rod opsin antibodies, Jeremy Nathans for cone opsin antibodies, Gabriel Travis for peripherin/RDS antibodies, and Thaddeus Dryja for discussions.